Magic-Mirror/gecko-dev
1
0
Fork 0
mirror of https://github.com/mozilla/gecko-dev.git synced 2025-02-02 03:49:37 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity

Bug 1309195 - Mark strBuf as empty after the contents have been used or ignored. r=wchen.

Browse Source 
MozReview-Commit-ID: 5KU62o2IMPs
This commit is contained in:
Henri Sivonen 2016-10-19 10:38:20 +03:00
parent b91bc4c721
commit e5ac89f389
3 changed files with 354 additions and 256 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

171
parser/html/javasrc/Tokenizer.java
View File

@ -35,6 +35,11 @@
package nu.validator.htmlparser.impl;
import org.xml.sax.ErrorHandler;
import org.xml.sax.Locator;
import org.xml.sax.SAXException;
import org.xml.sax.SAXParseException;
import nu.validator.htmlparser.annotation.Auto;
import nu.validator.htmlparser.annotation.CharacterName;
import nu.validator.htmlparser.annotation.Const;
@ -46,11 +51,6 @@ import nu.validator.htmlparser.common.Interner;
import nu.validator.htmlparser.common.TokenHandler;
import nu.validator.htmlparser.common.XmlViolationPolicy;
import org.xml.sax.ErrorHandler;
import org.xml.sax.Locator;
import org.xml.sax.SAXException;
import org.xml.sax.SAXParseException;
/**
* An implementation of
* https://html.spec.whatwg.org/multipage/syntax.html#tokenization
@ -822,27 +822,29 @@ public class Tokenizer implements Locator {
charRefBuf[charRefBufLen++] = c;
}
@Inline private void clearCharRefBufAndAppend(char c) {
charRefBuf[0] = c;
charRefBufLen = 1;
}
private void emitOrAppendCharRefBuf(int returnState) throws SAXException {
if ((returnState & DATA_AND_RCDATA_MASK) != 0) {
appendCharRefBufToStrBuf();
} else {
if (charRefBufLen > 0) {
tokenHandler.characters(charRefBuf, 0, charRefBufLen);
charRefBufLen = 0;
}
}
}
@Inline private void clearStrBufAndAppend(char c) {
strBuf[0] = c;
strBufLen = 1;
@Inline private void clearStrBufAfterUse() {
strBufLen = 0;
}
@Inline private void clearStrBuf() {
@Inline private void clearStrBufBeforeUse() {
assert strBufLen == 0: "strBufLen not reset after previous use!";
strBufLen = 0; // no-op in the absence of bugs
}
@Inline private void clearStrBufAfterOneHyphen() {
assert strBufLen == 1: "strBufLen length not one!";
assert strBuf[0] == '-': "strBuf does not start with a hyphen!";
strBufLen = 0;
}
@ -871,9 +873,11 @@ public class Tokenizer implements Locator {
* @return the buffer as a string
*/
protected String strBufToString() {
return Portability.newStringFromBuffer(strBuf, 0, strBufLen
String str = Portability.newStringFromBuffer(strBuf, 0, strBufLen
// CPPONLY: , tokenHandler
);
clearStrBufAfterUse();
return str;
}
/**
@ -885,6 +889,7 @@ public class Tokenizer implements Locator {
private void strBufToDoctypeName() {
doctypeName = Portability.newLocalNameFromBuffer(strBuf, 0, strBufLen,
interner);
clearStrBufAfterUse();
}
/**
@ -896,6 +901,7 @@ public class Tokenizer implements Locator {
private void emitStrBuf() throws SAXException {
if (strBufLen > 0) {
tokenHandler.characters(strBuf, 0, strBufLen);
clearStrBufAfterUse();
}
}
@ -942,6 +948,8 @@ public class Tokenizer implements Locator {
switch (commentPolicy) {
case ALTER_INFOSET:
strBufLen--;
// WARNING!!! This expands the worst case of the buffer length
// given the length of input!
appendStrBuf(' ');
appendStrBuf('-');
// FALLTHROUGH
@ -975,6 +983,7 @@ public class Tokenizer implements Locator {
*/
@Inline private void appendCharRefBufToStrBuf() {
appendStrBuf(charRefBuf, 0, charRefBufLen);
charRefBufLen = 0;
}
/**
@ -995,6 +1004,7 @@ public class Tokenizer implements Locator {
// [NOCPP[
}
// ]NOCPP]
clearStrBufAfterUse();
cstart = pos + 1;
}
@ -1082,6 +1092,7 @@ public class Tokenizer implements Locator {
private void strBufToElementNameString() {
tagName = ElementName.elementNameByBuffer(strBuf, 0, strBufLen,
interner);
clearStrBufAfterUse();
}
private int emitCurrentTagToken(boolean selfClosing, int pos)
@ -1133,6 +1144,7 @@ public class Tokenizer implements Locator {
, namePolicy != XmlViolationPolicy.ALLOW
// ]NOCPP]
, interner);
clearStrBufAfterUse();
if (attributes == null) {
attributes = new HtmlAttributes(mappingLangToXmlLang);
@ -1199,6 +1211,8 @@ public class Tokenizer implements Locator {
// ]NOCPP]
attributeName = null; // attributeName has been adopted by the
// |attributes| object
} else {
clearStrBufAfterUse();
}
}
@ -1229,6 +1243,9 @@ public class Tokenizer implements Locator {
);
attributeName = null; // attributeName has been adopted by the
// |attributes| object
} else {
// We have a duplicate attribute. Explicitly discard its value.
clearStrBufAfterUse();
}
}
@ -1347,6 +1364,14 @@ public class Tokenizer implements Locator {
// unifying the tokenizer and tree builder buffers in the future.
int worstCase = strBufLen + inputLength + charRefBufLen + 2;
tokenHandler.ensureBufferSpace(worstCase);
if (commentPolicy == XmlViolationPolicy.ALTER_INFOSET) {
// When altering infoset, if the comment contents are consecutive
// hyphens, each hyphen generates a space, too. These buffer
// contents never get emitted as characters() to the tokenHandler,
// which is why this calculation happens after the call to
// ensureBufferSpace on tokenHandler.
worstCase *= 2;
}
if (strBuf == null) {
// Add an arbitrary small value to avoid immediate reallocation
// once there are a few characters in the buffer.
@ -1455,7 +1480,8 @@ public class Tokenizer implements Locator {
* reference in data state.
*/
flushChars(buf, pos);
clearCharRefBufAndAppend(c);
assert charRefBufLen == 0: "charRefBufLen not reset after previous use!";
appendCharRefBuf(c);
setAdditionalAndRememberAmpersandLocation('\u0000');
returnState = state;
state = transition(state, Tokenizer.CONSUME_CHARACTER_REFERENCE, reconsume, pos);
@ -1515,7 +1541,8 @@ public class Tokenizer implements Locator {
* input character (add 0x0020 to the character's
* code point),
*/
clearStrBufAndAppend((char) (c + 0x20));
clearStrBufBeforeUse();
appendStrBuf((char) (c + 0x20));
/* then switch to the tag name state. */
state = transition(state, Tokenizer.TAG_NAME, reconsume, pos);
/*
@ -1534,7 +1561,8 @@ public class Tokenizer implements Locator {
/*
* set its tag name to the input character,
*/
clearStrBufAndAppend(c);
clearStrBufBeforeUse();
appendStrBuf(c);
/* then switch to the tag name state. */
state = transition(state, Tokenizer.TAG_NAME, reconsume, pos);
/*
@ -1574,7 +1602,8 @@ public class Tokenizer implements Locator {
/*
* Switch to the bogus comment state.
*/
clearStrBufAndAppend(c);
clearStrBufBeforeUse();
appendStrBuf(c);
state = transition(state, Tokenizer.BOGUS_COMMENT, reconsume, pos);
continue stateloop;
case '>':
@ -1777,7 +1806,8 @@ public class Tokenizer implements Locator {
* Set that attribute's name to the current
* input character,
*/
clearStrBufAndAppend(c);
clearStrBufBeforeUse();
appendStrBuf(c);
/*
* and its value to the empty string.
*/
@ -1923,7 +1953,7 @@ public class Tokenizer implements Locator {
* attribute value (double-quoted) state.
*/
// CPPONLY: attributeLine = line;
clearStrBuf();
clearStrBufBeforeUse();
state = transition(state, Tokenizer.ATTRIBUTE_VALUE_DOUBLE_QUOTED, reconsume, pos);
break beforeattributevalueloop;
// continue stateloop;
@ -1934,7 +1964,7 @@ public class Tokenizer implements Locator {
* input character.
*/
// CPPONLY: attributeLine = line;
clearStrBuf();
clearStrBufBeforeUse();
reconsume = true;
state = transition(state, Tokenizer.ATTRIBUTE_VALUE_UNQUOTED, reconsume, pos);
noteUnquotedAttributeValue();
@ -1945,7 +1975,7 @@ public class Tokenizer implements Locator {
* value (single-quoted) state.
*/
// CPPONLY: attributeLine = line;
clearStrBuf();
clearStrBufBeforeUse();
state = transition(state, Tokenizer.ATTRIBUTE_VALUE_SINGLE_QUOTED, reconsume, pos);
continue stateloop;
case '>':
@ -1989,7 +2019,8 @@ public class Tokenizer implements Locator {
* character to the current attribute's value.
*/
// CPPONLY: attributeLine = line;
clearStrBufAndAppend(c);
clearStrBufBeforeUse();
appendStrBuf(c);
/*
* Switch to the attribute value (unquoted)
* state.
@ -2032,7 +2063,8 @@ public class Tokenizer implements Locator {
* additional allowed character being U+0022
* QUOTATION MARK (").
*/
clearCharRefBufAndAppend(c);
assert charRefBufLen == 0: "charRefBufLen not reset after previous use!";
appendCharRefBuf(c);
setAdditionalAndRememberAmpersandLocation('\"');
returnState = state;
state = transition(state, Tokenizer.CONSUME_CHARACTER_REFERENCE, reconsume, pos);
@ -2201,7 +2233,8 @@ public class Tokenizer implements Locator {
* additional allowed character being U+003E
* GREATER-THAN SIGN (>)
*/
clearCharRefBufAndAppend(c);
assert charRefBufLen == 0: "charRefBufLen not reset after previous use!";
appendCharRefBuf(c);
setAdditionalAndRememberAmpersandLocation('>');
returnState = state;
state = transition(state, Tokenizer.CONSUME_CHARACTER_REFERENCE, reconsume, pos);
@ -2340,7 +2373,8 @@ public class Tokenizer implements Locator {
* Set that attribute's name to the current
* input character,
*/
clearStrBufAndAppend(c);
clearStrBufBeforeUse();
appendStrBuf(c);
/*
* and its value to the empty string.
*/
@ -2386,19 +2420,22 @@ public class Tokenizer implements Locator {
*/
switch (c) {
case '-':
clearStrBufAndAppend(c);
clearStrBufBeforeUse();
appendStrBuf(c);
state = transition(state, Tokenizer.MARKUP_DECLARATION_HYPHEN, reconsume, pos);
break markupdeclarationopenloop;
// continue stateloop;
case 'd':
case 'D':
clearStrBufAndAppend(c);
clearStrBufBeforeUse();
appendStrBuf(c);
index = 0;
state = transition(state, Tokenizer.MARKUP_DECLARATION_OCTYPE, reconsume, pos);
continue stateloop;
case '[':
if (tokenHandler.cdataSectionAllowed()) {
clearStrBufAndAppend(c);
clearStrBufBeforeUse();
appendStrBuf(c);
index = 0;
state = transition(state, Tokenizer.CDATA_START, reconsume, pos);
continue stateloop;
@ -2406,7 +2443,7 @@ public class Tokenizer implements Locator {
// else fall through
default:
errBogusComment();
clearStrBuf();
clearStrBufBeforeUse();
reconsume = true;
state = transition(state, Tokenizer.BOGUS_COMMENT, reconsume, pos);
continue stateloop;
@ -2423,7 +2460,7 @@ public class Tokenizer implements Locator {
case '\u0000':
break stateloop;
case '-':
clearStrBuf();
clearStrBufAfterOneHyphen();
state = transition(state, Tokenizer.COMMENT_START, reconsume, pos);
break markupdeclarationhyphenloop;
// continue stateloop;
@ -2777,6 +2814,7 @@ public class Tokenizer implements Locator {
index++;
continue;
} else {
clearStrBufAfterUse();
cstart = pos; // start coalescing
reconsume = true;
state = transition(state, Tokenizer.CDATA_SECTION, reconsume, pos);
@ -2890,7 +2928,8 @@ public class Tokenizer implements Locator {
* + additional allowed character being U+0027
* APOSTROPHE (').
*/
clearCharRefBufAndAppend(c);
assert charRefBufLen == 0: "charRefBufLen not reset after previous use!";
appendCharRefBuf(c);
setAdditionalAndRememberAmpersandLocation('\'');
returnState = state;
state = transition(state, Tokenizer.CONSUME_CHARACTER_REFERENCE, reconsume, pos);
@ -3238,6 +3277,8 @@ public class Tokenizer implements Locator {
charRefBufLen - charRefBufMark);
}
}
// charRefBufLen will be zeroed below!
// Check if we broke out early with c being the last
// character that matched as opposed to being the
// first one that didn't match. In the case of an
@ -3245,6 +3286,7 @@ public class Tokenizer implements Locator {
// *after* the current character and the current
// character shouldn't be reconsumed.
boolean earlyBreak = (c == ';' && charRefBufMark == charRefBufLen);
charRefBufLen = 0;
if ((returnState & DATA_AND_RCDATA_MASK) == 0) {
cstart = earlyBreak ? pos + 1 : pos;
}
@ -3383,6 +3425,8 @@ public class Tokenizer implements Locator {
// WARNING FALLTHRU CASE TRANSITION: DON'T REORDER
case HANDLE_NCR_VALUE:
// WARNING previous state sets reconsume
// We are not going to emit the contents of charRefBuf.
charRefBufLen = 0;
// XXX inline this case if the method size can take it
handleNcrValue(returnState);
state = transition(state, returnState, reconsume, pos);
@ -3529,7 +3573,8 @@ public class Tokenizer implements Locator {
/*
* Switch to the bogus comment state.
*/
clearStrBufAndAppend('\n');
clearStrBufBeforeUse();
appendStrBuf('\n');
state = transition(state, Tokenizer.BOGUS_COMMENT, reconsume, pos);
break stateloop;
case '\n':
@ -3539,7 +3584,8 @@ public class Tokenizer implements Locator {
/*
* Switch to the bogus comment state.
*/
clearStrBufAndAppend('\n');
clearStrBufBeforeUse();
appendStrBuf(c);
state = transition(state, Tokenizer.BOGUS_COMMENT, reconsume, pos);
continue stateloop;
case '\u0000':
@ -3559,7 +3605,8 @@ public class Tokenizer implements Locator {
/*
* set its tag name to the input character,
*/
clearStrBufAndAppend(c);
clearStrBufBeforeUse();
appendStrBuf(c);
/*
* then switch to the tag name state. (Don't
* emit the token yet; further details will be
@ -3573,7 +3620,8 @@ public class Tokenizer implements Locator {
/*
* Switch to the bogus comment state.
*/
clearStrBufAndAppend(c);
clearStrBufBeforeUse();
appendStrBuf(c);
state = transition(state, Tokenizer.BOGUS_COMMENT, reconsume, pos);
continue stateloop;
}
@ -3596,7 +3644,8 @@ public class Tokenizer implements Locator {
* reference in RCDATA state.
*/
flushChars(buf, pos);
clearCharRefBufAndAppend(c);
assert charRefBufLen == 0: "charRefBufLen not reset after previous use!";
appendCharRefBuf(c);
setAdditionalAndRememberAmpersandLocation('\u0000');
returnState = state;
state = transition(state, Tokenizer.CONSUME_CHARACTER_REFERENCE, reconsume, pos);
@ -3681,7 +3730,7 @@ public class Tokenizer implements Locator {
* data end tag open state.
*/
index = 0;
clearStrBuf();
clearStrBufBeforeUse();
state = transition(state, Tokenizer.NON_DATA_END_TAG_NAME, reconsume, pos);
break rawtextrcdatalessthansignloop;
// FALL THRU continue stateloop;
@ -3710,7 +3759,7 @@ public class Tokenizer implements Locator {
c = checkChar(buf, pos);
/*
* ASSERT! when entering this state, set index to 0 and
* call clearStrBuf() assert (contentModelElement !=
* call clearStrBufBeforeUse() assert (contentModelElement !=
* null); Let's implement the above without lookahead.
* strBuf is the 'temporary buffer'.
*/
@ -3743,6 +3792,7 @@ public class Tokenizer implements Locator {
switch (c) {
case '\r':
silentCarriageReturn();
clearStrBufAfterUse(); // strBuf not used
state = transition(state, Tokenizer.BEFORE_ATTRIBUTE_NAME, reconsume, pos);
break stateloop;
case '\n':
@ -3758,6 +3808,7 @@ public class Tokenizer implements Locator {
* appropriate end tag token, then switch to
* the before attribute name state.
*/
clearStrBufAfterUse(); // strBuf not used
state = transition(state, Tokenizer.BEFORE_ATTRIBUTE_NAME, reconsume, pos);
continue stateloop;
case '/':
@ -3767,6 +3818,7 @@ public class Tokenizer implements Locator {
* then switch to the self-closing start tag
* state.
*/
clearStrBufAfterUse(); // strBuf not used
state = transition(state, Tokenizer.SELF_CLOSING_START_TAG, reconsume, pos);
continue stateloop;
case '>':
@ -3776,6 +3828,7 @@ public class Tokenizer implements Locator {
* end tag token, then emit the current tag
* token and switch to the data state.
*/
clearStrBufAfterUse(); // strBuf not used
state = transition(state, emitCurrentTagToken(false, pos), reconsume, pos);
if (shouldSuspend) {
break stateloop;
@ -3950,7 +4003,7 @@ public class Tokenizer implements Locator {
* data end tag open state.
*/
index = 0;
clearStrBuf();
clearStrBufBeforeUse();
state = transition(state, Tokenizer.NON_DATA_END_TAG_NAME, reconsume, pos);
continue stateloop;
case '!':
@ -4208,7 +4261,7 @@ public class Tokenizer implements Locator {
* data escaped end tag open state.
*/
index = 0;
clearStrBuf();
clearStrBufBeforeUse();
returnState = Tokenizer.SCRIPT_DATA_ESCAPED;
state = transition(state, Tokenizer.NON_DATA_END_TAG_NAME, reconsume, pos);
continue stateloop;
@ -4690,7 +4743,8 @@ public class Tokenizer implements Locator {
* Set the token's name name to the current
* input character.
*/
clearStrBufAndAppend(c);
clearStrBufBeforeUse();
appendStrBuf(c);
/*
* Switch to the DOCTYPE name state.
*/
@ -4911,7 +4965,7 @@ public class Tokenizer implements Locator {
* Set the DOCTYPE token's public identifier to
* the empty string (not missing),
*/
clearStrBuf();
clearStrBufBeforeUse();
/*
* then switch to the DOCTYPE public identifier
* (double-quoted) state.
@ -4927,7 +4981,7 @@ public class Tokenizer implements Locator {
* Set the DOCTYPE token's public identifier to
* the empty string (not missing),
*/
clearStrBuf();
clearStrBufBeforeUse();
/*
* then switch to the DOCTYPE public identifier
* (single-quoted) state.
@ -4998,7 +5052,7 @@ public class Tokenizer implements Locator {
* token's public identifier to the empty string
* (not missing),
*/
clearStrBuf();
clearStrBufBeforeUse();
/*
* then switch to the DOCTYPE public identifier
* (double-quoted) state.
@ -5012,7 +5066,7 @@ public class Tokenizer implements Locator {
* public identifier to the empty string (not
* missing),
*/
clearStrBuf();
clearStrBufBeforeUse();
/*
* then switch to the DOCTYPE public identifier
* (single-quoted) state.
@ -5163,7 +5217,7 @@ public class Tokenizer implements Locator {
* Set the DOCTYPE token's system identifier to
* the empty string (not missing),
*/
clearStrBuf();
clearStrBufBeforeUse();
/*
* then switch to the DOCTYPE system identifier
* (double-quoted) state.
@ -5179,7 +5233,7 @@ public class Tokenizer implements Locator {
* Set the DOCTYPE token's system identifier to
* the empty string (not missing),
*/
clearStrBuf();
clearStrBufBeforeUse();
/*
* then switch to the DOCTYPE system identifier
* (single-quoted) state.
@ -5244,7 +5298,7 @@ public class Tokenizer implements Locator {
* token's system identifier to the empty string
* (not missing),
*/
clearStrBuf();
clearStrBufBeforeUse();
/*
* then switch to the DOCTYPE system identifier
* (double-quoted) state.
@ -5258,7 +5312,7 @@ public class Tokenizer implements Locator {
* system identifier to the empty string (not
* missing),
*/
clearStrBuf();
clearStrBufBeforeUse();
/*
* then switch to the DOCTYPE system identifier
* (single-quoted) state.
@ -5507,7 +5561,7 @@ public class Tokenizer implements Locator {
* Set the DOCTYPE token's system identifier to
* the empty string (not missing),
*/
clearStrBuf();
clearStrBufBeforeUse();
/*
* then switch to the DOCTYPE public identifier
* (double-quoted) state.
@ -5523,7 +5577,7 @@ public class Tokenizer implements Locator {
* Set the DOCTYPE token's public identifier to
* the empty string (not missing),
*/
clearStrBuf();
clearStrBufBeforeUse();
/*
* then switch to the DOCTYPE public identifier
* (single-quoted) state.
@ -5594,7 +5648,7 @@ public class Tokenizer implements Locator {
* token's system identifier to the empty string
* (not missing),
*/
clearStrBuf();
clearStrBufBeforeUse();
/*
* then switch to the DOCTYPE system identifier
* (double-quoted) state.
@ -5607,7 +5661,7 @@ public class Tokenizer implements Locator {
* system identifier to the empty string (not
* missing),
*/
clearStrBuf();
clearStrBufBeforeUse();
/*
* then switch to the DOCTYPE system identifier
* (single-quoted) state.
@ -5823,6 +5877,9 @@ public class Tokenizer implements Locator {
// ]NOCPP]
private void initDoctypeFields() {
// Discard the characters "DOCTYPE" accumulated as a potential bogus
// comment into strBuf.
clearStrBufAfterUse();
doctypeName = "";
if (systemIdentifier != null) {
Portability.releaseString(systemIdentifier);
@ -6110,7 +6167,6 @@ public class Tokenizer implements Locator {
break eofloop;
case MARKUP_DECLARATION_OPEN:
errBogusComment();
clearStrBuf();
emitComment(0, 0);
break eofloop;
case MARKUP_DECLARATION_HYPHEN:
@ -6466,6 +6522,7 @@ public class Tokenizer implements Locator {
charRefBufLen - charRefBufMark);
}
}
charRefBufLen = 0;
state = returnState;
continue eofloop;
/*
@ -6645,7 +6702,7 @@ public class Tokenizer implements Locator {
}
public void resetToDataState() {
strBufLen = 0;
clearStrBufAfterUse();
charRefBufLen = 0;
stateSave = Tokenizer.DATA;
// line = 1; XXX line numbers

123
parser/html/nsHtml5Tokenizer.cpp
View File

@ -217,6 +217,7 @@ nsHtml5Tokenizer::emitOrAppendCharRefBuf(int32_t returnState)
} else {
if (charRefBufLen > 0) {
tokenHandler->characters(charRefBuf, 0, charRefBufLen);
charRefBufLen = 0;
}
}
}
@ -224,13 +225,16 @@ nsHtml5Tokenizer::emitOrAppendCharRefBuf(int32_t returnState)
nsString*
nsHtml5Tokenizer::strBufToString()
{
return nsHtml5Portability::newStringFromBuffer(strBuf, 0, strBufLen, tokenHandler);
nsString* str = nsHtml5Portability::newStringFromBuffer(strBuf, 0, strBufLen, tokenHandler);
clearStrBufAfterUse();
return str;
}
void
nsHtml5Tokenizer::strBufToDoctypeName()
{
doctypeName = nsHtml5Portability::newLocalNameFromBuffer(strBuf, 0, strBufLen, interner);
clearStrBufAfterUse();
}
void
@ -238,6 +242,7 @@ nsHtml5Tokenizer::emitStrBuf()
{
if (strBufLen > 0) {
tokenHandler->characters(strBuf, 0, strBufLen);
clearStrBufAfterUse();
}
}
@ -259,6 +264,7 @@ void
nsHtml5Tokenizer::emitComment(int32_t provisionalHyphens, int32_t pos)
{
tokenHandler->comment(strBuf, 0, strBufLen - provisionalHyphens);
clearStrBufAfterUse();
cstart = pos + 1;
}
@ -275,6 +281,7 @@ void
nsHtml5Tokenizer::strBufToElementNameString()
{
tagName = nsHtml5ElementName::elementNameByBuffer(strBuf, 0, strBufLen, interner);
clearStrBufAfterUse();
}
int32_t
@ -316,6 +323,7 @@ void
nsHtml5Tokenizer::attributeNameComplete()
{
attributeName = nsHtml5AttributeName::nameByBuffer(strBuf, 0, strBufLen, interner);
clearStrBufAfterUse();
if (!attributes) {
attributes = new nsHtml5HtmlAttributes(0);
}
@ -333,6 +341,8 @@ nsHtml5Tokenizer::addAttributeWithoutValue()
if (attributeName) {
attributes->addAttribute(attributeName, nsHtml5Portability::newEmptyString(), attributeLine);
attributeName = nullptr;
} else {
clearStrBufAfterUse();
}
}
@ -346,6 +356,8 @@ nsHtml5Tokenizer::addAttributeWithValue()
}
attributes->addAttribute(attributeName, val, attributeLine);
attributeName = nullptr;
} else {
clearStrBufAfterUse();
}
}
@ -427,7 +439,8 @@ nsHtml5Tokenizer::stateLoop(int32_t state, char16_t c, int32_t pos, char16_t* bu
switch(c) {
case '&': {
flushChars(buf, pos);
clearCharRefBufAndAppend(c);
MOZ_ASSERT(!charRefBufLen, "charRefBufLen not reset after previous use!");
appendCharRefBuf(c);
setAdditionalAndRememberAmpersandLocation('\0');
returnState = state;
state = P::transition(mViewSource, NS_HTML5TOKENIZER_CONSUME_CHARACTER_REFERENCE, reconsume, pos);
@ -464,12 +477,14 @@ nsHtml5Tokenizer::stateLoop(int32_t state, char16_t c, int32_t pos, char16_t* bu
c = checkChar(buf, pos);
if (c >= 'A' && c <= 'Z') {
endTag = false;
clearStrBufAndAppend((char16_t) (c + 0x20));
clearStrBufBeforeUse();
appendStrBuf((char16_t) (c + 0x20));
state = P::transition(mViewSource, NS_HTML5TOKENIZER_TAG_NAME, reconsume, pos);
NS_HTML5_BREAK(tagopenloop);
} else if (c >= 'a' && c <= 'z') {
endTag = false;
clearStrBufAndAppend(c);
clearStrBufBeforeUse();
appendStrBuf(c);
state = P::transition(mViewSource, NS_HTML5TOKENIZER_TAG_NAME, reconsume, pos);
NS_HTML5_BREAK(tagopenloop);
}
@ -490,7 +505,8 @@ nsHtml5Tokenizer::stateLoop(int32_t state, char16_t c, int32_t pos, char16_t* bu
if (P::reportErrors) {
errProcessingInstruction();
}
clearStrBufAndAppend(c);
clearStrBufBeforeUse();
appendStrBuf(c);
state = P::transition(mViewSource, NS_HTML5TOKENIZER_BOGUS_COMMENT, reconsume, pos);
NS_HTML5_CONTINUE(stateloop);
}
@ -617,7 +633,8 @@ nsHtml5Tokenizer::stateLoop(int32_t state, char16_t c, int32_t pos, char16_t* bu
c += 0x20;
}
attributeLine = line;
clearStrBufAndAppend(c);
clearStrBufBeforeUse();
appendStrBuf(c);
state = P::transition(mViewSource, NS_HTML5TOKENIZER_ATTRIBUTE_NAME, reconsume, pos);
NS_HTML5_BREAK(beforeattributenameloop);
}
@ -710,13 +727,13 @@ nsHtml5Tokenizer::stateLoop(int32_t state, char16_t c, int32_t pos, char16_t* bu
}
case '\"': {
attributeLine = line;
clearStrBuf();
clearStrBufBeforeUse();
state = P::transition(mViewSource, NS_HTML5TOKENIZER_ATTRIBUTE_VALUE_DOUBLE_QUOTED, reconsume, pos);
NS_HTML5_BREAK(beforeattributevalueloop);
}
case '&': {
attributeLine = line;
clearStrBuf();
clearStrBufBeforeUse();
reconsume = true;
state = P::transition(mViewSource, NS_HTML5TOKENIZER_ATTRIBUTE_VALUE_UNQUOTED, reconsume, pos);
@ -724,7 +741,7 @@ nsHtml5Tokenizer::stateLoop(int32_t state, char16_t c, int32_t pos, char16_t* bu
}
case '\'': {
attributeLine = line;
clearStrBuf();
clearStrBufBeforeUse();
state = P::transition(mViewSource, NS_HTML5TOKENIZER_ATTRIBUTE_VALUE_SINGLE_QUOTED, reconsume, pos);
NS_HTML5_CONTINUE(stateloop);
}
@ -751,7 +768,8 @@ nsHtml5Tokenizer::stateLoop(int32_t state, char16_t c, int32_t pos, char16_t* bu
}
default: {
attributeLine = line;
clearStrBufAndAppend(c);
clearStrBufBeforeUse();
appendStrBuf(c);
state = P::transition(mViewSource, NS_HTML5TOKENIZER_ATTRIBUTE_VALUE_UNQUOTED, reconsume, pos);
NS_HTML5_CONTINUE(stateloop);
@ -777,7 +795,8 @@ nsHtml5Tokenizer::stateLoop(int32_t state, char16_t c, int32_t pos, char16_t* bu
NS_HTML5_BREAK(attributevaluedoublequotedloop);
}
case '&': {
clearCharRefBufAndAppend(c);
MOZ_ASSERT(!charRefBufLen, "charRefBufLen not reset after previous use!");
appendCharRefBuf(c);
setAdditionalAndRememberAmpersandLocation('\"');
returnState = state;
state = P::transition(mViewSource, NS_HTML5TOKENIZER_CONSUME_CHARACTER_REFERENCE, reconsume, pos);
@ -897,7 +916,8 @@ nsHtml5Tokenizer::stateLoop(int32_t state, char16_t c, int32_t pos, char16_t* bu
NS_HTML5_CONTINUE(stateloop);
}
case '&': {
clearCharRefBufAndAppend(c);
MOZ_ASSERT(!charRefBufLen, "charRefBufLen not reset after previous use!");
appendCharRefBuf(c);
setAdditionalAndRememberAmpersandLocation('>');
returnState = state;
state = P::transition(mViewSource, NS_HTML5TOKENIZER_CONSUME_CHARACTER_REFERENCE, reconsume, pos);
@ -982,7 +1002,8 @@ nsHtml5Tokenizer::stateLoop(int32_t state, char16_t c, int32_t pos, char16_t* bu
if (c >= 'A' && c <= 'Z') {
c += 0x20;
}
clearStrBufAndAppend(c);
clearStrBufBeforeUse();
appendStrBuf(c);
state = P::transition(mViewSource, NS_HTML5TOKENIZER_ATTRIBUTE_NAME, reconsume, pos);
NS_HTML5_CONTINUE(stateloop);
}
@ -997,20 +1018,23 @@ nsHtml5Tokenizer::stateLoop(int32_t state, char16_t c, int32_t pos, char16_t* bu
c = checkChar(buf, pos);
switch(c) {
case '-': {
clearStrBufAndAppend(c);
clearStrBufBeforeUse();
appendStrBuf(c);
state = P::transition(mViewSource, NS_HTML5TOKENIZER_MARKUP_DECLARATION_HYPHEN, reconsume, pos);
NS_HTML5_BREAK(markupdeclarationopenloop);
}
case 'd':
case 'D': {
clearStrBufAndAppend(c);
clearStrBufBeforeUse();
appendStrBuf(c);
index = 0;
state = P::transition(mViewSource, NS_HTML5TOKENIZER_MARKUP_DECLARATION_OCTYPE, reconsume, pos);
NS_HTML5_CONTINUE(stateloop);
}
case '[': {
if (tokenHandler->cdataSectionAllowed()) {
clearStrBufAndAppend(c);
clearStrBufBeforeUse();
appendStrBuf(c);
index = 0;
state = P::transition(mViewSource, NS_HTML5TOKENIZER_CDATA_START, reconsume, pos);
NS_HTML5_CONTINUE(stateloop);
@ -1020,7 +1044,7 @@ nsHtml5Tokenizer::stateLoop(int32_t state, char16_t c, int32_t pos, char16_t* bu
if (P::reportErrors) {
errBogusComment();
}
clearStrBuf();
clearStrBufBeforeUse();
reconsume = true;
state = P::transition(mViewSource, NS_HTML5TOKENIZER_BOGUS_COMMENT, reconsume, pos);
NS_HTML5_CONTINUE(stateloop);
@ -1040,7 +1064,7 @@ nsHtml5Tokenizer::stateLoop(int32_t state, char16_t c, int32_t pos, char16_t* bu
NS_HTML5_BREAK(stateloop);
}
case '-': {
clearStrBuf();
clearStrBufAfterOneHyphen();
state = P::transition(mViewSource, NS_HTML5TOKENIZER_COMMENT_START, reconsume, pos);
NS_HTML5_BREAK(markupdeclarationhyphenloop);
}
@ -1304,6 +1328,7 @@ nsHtml5Tokenizer::stateLoop(int32_t state, char16_t c, int32_t pos, char16_t* bu
index++;
continue;
} else {
clearStrBufAfterUse();
cstart = pos;
reconsume = true;
state = P::transition(mViewSource, NS_HTML5TOKENIZER_CDATA_SECTION, reconsume, pos);
@ -1411,7 +1436,8 @@ nsHtml5Tokenizer::stateLoop(int32_t state, char16_t c, int32_t pos, char16_t* bu
NS_HTML5_CONTINUE(stateloop);
}
case '&': {
clearCharRefBufAndAppend(c);
MOZ_ASSERT(!charRefBufLen, "charRefBufLen not reset after previous use!");
appendCharRefBuf(c);
setAdditionalAndRememberAmpersandLocation('\'');
returnState = state;
state = P::transition(mViewSource, NS_HTML5TOKENIZER_CONSUME_CHARACTER_REFERENCE, reconsume, pos);
@ -1643,6 +1669,7 @@ nsHtml5Tokenizer::stateLoop(int32_t state, char16_t c, int32_t pos, char16_t* bu
}
}
bool earlyBreak = (c == ';' && charRefBufMark == charRefBufLen);
charRefBufLen = 0;
if (!(returnState & NS_HTML5TOKENIZER_DATA_AND_RCDATA_MASK)) {
cstart = earlyBreak ? pos + 1 : pos;
}
@ -1736,6 +1763,7 @@ nsHtml5Tokenizer::stateLoop(int32_t state, char16_t c, int32_t pos, char16_t* bu
decimalloop_end: ;
}
case NS_HTML5TOKENIZER_HANDLE_NCR_VALUE: {
charRefBufLen = 0;
handleNcrValue(returnState);
state = P::transition(mViewSource, returnState, reconsume, pos);
NS_HTML5_CONTINUE(stateloop);
@ -1861,7 +1889,8 @@ nsHtml5Tokenizer::stateLoop(int32_t state, char16_t c, int32_t pos, char16_t* bu
if (P::reportErrors) {
errGarbageAfterLtSlash();
}
clearStrBufAndAppend('\n');
clearStrBufBeforeUse();
appendStrBuf('\n');
state = P::transition(mViewSource, NS_HTML5TOKENIZER_BOGUS_COMMENT, reconsume, pos);
NS_HTML5_BREAK(stateloop);
}
@ -1870,7 +1899,8 @@ nsHtml5Tokenizer::stateLoop(int32_t state, char16_t c, int32_t pos, char16_t* bu
if (P::reportErrors) {
errGarbageAfterLtSlash();
}
clearStrBufAndAppend('\n');
clearStrBufBeforeUse();
appendStrBuf(c);
state = P::transition(mViewSource, NS_HTML5TOKENIZER_BOGUS_COMMENT, reconsume, pos);
NS_HTML5_CONTINUE(stateloop);
}
@ -1883,14 +1913,16 @@ nsHtml5Tokenizer::stateLoop(int32_t state, char16_t c, int32_t pos, char16_t* bu
}
if (c >= 'a' && c <= 'z') {
endTag = true;
clearStrBufAndAppend(c);
clearStrBufBeforeUse();
appendStrBuf(c);
state = P::transition(mViewSource, NS_HTML5TOKENIZER_TAG_NAME, reconsume, pos);
NS_HTML5_CONTINUE(stateloop);
} else {
if (P::reportErrors) {
errGarbageAfterLtSlash();
}
clearStrBufAndAppend(c);
clearStrBufBeforeUse();
appendStrBuf(c);
state = P::transition(mViewSource, NS_HTML5TOKENIZER_BOGUS_COMMENT, reconsume, pos);
NS_HTML5_CONTINUE(stateloop);
}
@ -1910,7 +1942,8 @@ nsHtml5Tokenizer::stateLoop(int32_t state, char16_t c, int32_t pos, char16_t* bu
switch(c) {
case '&': {
flushChars(buf, pos);
clearCharRefBufAndAppend(c);
MOZ_ASSERT(!charRefBufLen, "charRefBufLen not reset after previous use!");
appendCharRefBuf(c);
setAdditionalAndRememberAmpersandLocation('\0');
returnState = state;
state = P::transition(mViewSource, NS_HTML5TOKENIZER_CONSUME_CHARACTER_REFERENCE, reconsume, pos);
@ -1984,7 +2017,7 @@ nsHtml5Tokenizer::stateLoop(int32_t state, char16_t c, int32_t pos, char16_t* bu
switch(c) {
case '/': {
index = 0;
clearStrBuf();
clearStrBufBeforeUse();
state = P::transition(mViewSource, NS_HTML5TOKENIZER_NON_DATA_END_TAG_NAME, reconsume, pos);
NS_HTML5_BREAK(rawtextrcdatalessthansignloop);
}
@ -2028,6 +2061,7 @@ nsHtml5Tokenizer::stateLoop(int32_t state, char16_t c, int32_t pos, char16_t* bu
switch(c) {
case '\r': {
silentCarriageReturn();
clearStrBufAfterUse();
state = P::transition(mViewSource, NS_HTML5TOKENIZER_BEFORE_ATTRIBUTE_NAME, reconsume, pos);
NS_HTML5_BREAK(stateloop);
}
@ -2037,14 +2071,17 @@ nsHtml5Tokenizer::stateLoop(int32_t state, char16_t c, int32_t pos, char16_t* bu
case ' ':
case '\t':
case '\f': {
clearStrBufAfterUse();
state = P::transition(mViewSource, NS_HTML5TOKENIZER_BEFORE_ATTRIBUTE_NAME, reconsume, pos);
NS_HTML5_CONTINUE(stateloop);
}
case '/': {
clearStrBufAfterUse();
state = P::transition(mViewSource, NS_HTML5TOKENIZER_SELF_CLOSING_START_TAG, reconsume, pos);
NS_HTML5_CONTINUE(stateloop);
}
case '>': {
clearStrBufAfterUse();
state = P::transition(mViewSource, emitCurrentTagToken(false, pos), reconsume, pos);
if (shouldSuspend) {
NS_HTML5_BREAK(stateloop);
@ -2188,7 +2225,7 @@ nsHtml5Tokenizer::stateLoop(int32_t state, char16_t c, int32_t pos, char16_t* bu
switch(c) {
case '/': {
index = 0;
clearStrBuf();
clearStrBufBeforeUse();
state = P::transition(mViewSource, NS_HTML5TOKENIZER_NON_DATA_END_TAG_NAME, reconsume, pos);
NS_HTML5_CONTINUE(stateloop);
}
@ -2373,7 +2410,7 @@ nsHtml5Tokenizer::stateLoop(int32_t state, char16_t c, int32_t pos, char16_t* bu
switch(c) {
case '/': {
index = 0;
clearStrBuf();
clearStrBufBeforeUse();
returnState = NS_HTML5TOKENIZER_SCRIPT_DATA_ESCAPED;
state = P::transition(mViewSource, NS_HTML5TOKENIZER_NON_DATA_END_TAG_NAME, reconsume, pos);
NS_HTML5_CONTINUE(stateloop);
@ -2729,7 +2766,8 @@ nsHtml5Tokenizer::stateLoop(int32_t state, char16_t c, int32_t pos, char16_t* bu
if (c >= 'A' && c <= 'Z') {
c += 0x20;
}
clearStrBufAndAppend(c);
clearStrBufBeforeUse();
appendStrBuf(c);
state = P::transition(mViewSource, NS_HTML5TOKENIZER_DOCTYPE_NAME, reconsume, pos);
NS_HTML5_BREAK(beforedoctypenameloop);
}
@ -2881,7 +2919,7 @@ nsHtml5Tokenizer::stateLoop(int32_t state, char16_t c, int32_t pos, char16_t* bu
if (P::reportErrors) {
errNoSpaceBetweenDoctypePublicKeywordAndQuote();
}
clearStrBuf();
clearStrBufBeforeUse();
state = P::transition(mViewSource, NS_HTML5TOKENIZER_DOCTYPE_PUBLIC_IDENTIFIER_DOUBLE_QUOTED, reconsume, pos);
NS_HTML5_CONTINUE(stateloop);
}
@ -2889,7 +2927,7 @@ nsHtml5Tokenizer::stateLoop(int32_t state, char16_t c, int32_t pos, char16_t* bu
if (P::reportErrors) {
errNoSpaceBetweenDoctypePublicKeywordAndQuote();
}
clearStrBuf();
clearStrBufBeforeUse();
state = P::transition(mViewSource, NS_HTML5TOKENIZER_DOCTYPE_PUBLIC_IDENTIFIER_SINGLE_QUOTED, reconsume, pos);
NS_HTML5_CONTINUE(stateloop);
}
@ -2931,12 +2969,12 @@ nsHtml5Tokenizer::stateLoop(int32_t state, char16_t c, int32_t pos, char16_t* bu
continue;
}
case '\"': {
clearStrBuf();
clearStrBufBeforeUse();
state = P::transition(mViewSource, NS_HTML5TOKENIZER_DOCTYPE_PUBLIC_IDENTIFIER_DOUBLE_QUOTED, reconsume, pos);
NS_HTML5_BREAK(beforedoctypepublicidentifierloop);
}
case '\'': {
clearStrBuf();
clearStrBufBeforeUse();
state = P::transition(mViewSource, NS_HTML5TOKENIZER_DOCTYPE_PUBLIC_IDENTIFIER_SINGLE_QUOTED, reconsume, pos);
NS_HTML5_CONTINUE(stateloop);
}
@ -3029,7 +3067,7 @@ nsHtml5Tokenizer::stateLoop(int32_t state, char16_t c, int32_t pos, char16_t* bu
if (P::reportErrors) {
errNoSpaceBetweenPublicAndSystemIds();
}
clearStrBuf();
clearStrBufBeforeUse();
state = P::transition(mViewSource, NS_HTML5TOKENIZER_DOCTYPE_SYSTEM_IDENTIFIER_DOUBLE_QUOTED, reconsume, pos);
NS_HTML5_CONTINUE(stateloop);
}
@ -3037,7 +3075,7 @@ nsHtml5Tokenizer::stateLoop(int32_t state, char16_t c, int32_t pos, char16_t* bu
if (P::reportErrors) {
errNoSpaceBetweenPublicAndSystemIds();
}
clearStrBuf();
clearStrBufBeforeUse();
state = P::transition(mViewSource, NS_HTML5TOKENIZER_DOCTYPE_SYSTEM_IDENTIFIER_SINGLE_QUOTED, reconsume, pos);
NS_HTML5_CONTINUE(stateloop);
}
@ -3075,12 +3113,12 @@ nsHtml5Tokenizer::stateLoop(int32_t state, char16_t c, int32_t pos, char16_t* bu
NS_HTML5_CONTINUE(stateloop);
}
case '\"': {
clearStrBuf();
clearStrBufBeforeUse();
state = P::transition(mViewSource, NS_HTML5TOKENIZER_DOCTYPE_SYSTEM_IDENTIFIER_DOUBLE_QUOTED, reconsume, pos);
NS_HTML5_BREAK(betweendoctypepublicandsystemidentifiersloop);
}
case '\'': {
clearStrBuf();
clearStrBufBeforeUse();
state = P::transition(mViewSource, NS_HTML5TOKENIZER_DOCTYPE_SYSTEM_IDENTIFIER_SINGLE_QUOTED, reconsume, pos);
NS_HTML5_CONTINUE(stateloop);
}
@ -3252,7 +3290,7 @@ nsHtml5Tokenizer::stateLoop(int32_t state, char16_t c, int32_t pos, char16_t* bu
if (P::reportErrors) {
errNoSpaceBetweenDoctypeSystemKeywordAndQuote();
}
clearStrBuf();
clearStrBufBeforeUse();
state = P::transition(mViewSource, NS_HTML5TOKENIZER_DOCTYPE_SYSTEM_IDENTIFIER_DOUBLE_QUOTED, reconsume, pos);
NS_HTML5_CONTINUE(stateloop);
}
@ -3260,7 +3298,7 @@ nsHtml5Tokenizer::stateLoop(int32_t state, char16_t c, int32_t pos, char16_t* bu
if (P::reportErrors) {
errNoSpaceBetweenDoctypeSystemKeywordAndQuote();
}
clearStrBuf();
clearStrBufBeforeUse();
state = P::transition(mViewSource, NS_HTML5TOKENIZER_DOCTYPE_SYSTEM_IDENTIFIER_SINGLE_QUOTED, reconsume, pos);
NS_HTML5_CONTINUE(stateloop);
}
@ -3302,12 +3340,12 @@ nsHtml5Tokenizer::stateLoop(int32_t state, char16_t c, int32_t pos, char16_t* bu
continue;
}
case '\"': {
clearStrBuf();
clearStrBufBeforeUse();
state = P::transition(mViewSource, NS_HTML5TOKENIZER_DOCTYPE_SYSTEM_IDENTIFIER_DOUBLE_QUOTED, reconsume, pos);
NS_HTML5_CONTINUE(stateloop);
}
case '\'': {
clearStrBuf();
clearStrBufBeforeUse();
state = P::transition(mViewSource, NS_HTML5TOKENIZER_DOCTYPE_SYSTEM_IDENTIFIER_SINGLE_QUOTED, reconsume, pos);
NS_HTML5_BREAK(beforedoctypesystemidentifierloop);
}
@ -3455,6 +3493,7 @@ nsHtml5Tokenizer::stateLoop(int32_t state, char16_t c, int32_t pos, char16_t* bu
void
nsHtml5Tokenizer::initDoctypeFields()
{
clearStrBufAfterUse();
doctypeName = nsHtml5Atoms::emptystring;
if (systemIdentifier) {
nsHtml5Portability::releaseString(systemIdentifier);
@ -3607,7 +3646,6 @@ nsHtml5Tokenizer::eof()
}
case NS_HTML5TOKENIZER_MARKUP_DECLARATION_OPEN: {
errBogusComment();
clearStrBuf();
emitComment(0, 0);
NS_HTML5_BREAK(eofloop);
}
@ -3813,6 +3851,7 @@ nsHtml5Tokenizer::eof()
tokenHandler->characters(charRefBuf, charRefBufMark, charRefBufLen - charRefBufMark);
}
}
charRefBufLen = 0;
state = returnState;
NS_HTML5_CONTINUE(eofloop);
}
@ -3935,7 +3974,7 @@ nsHtml5Tokenizer::isInDataState()
void
nsHtml5Tokenizer::resetToDataState()
{
strBufLen = 0;
clearStrBufAfterUse();
charRefBufLen = 0;
stateSave = NS_HTML5TOKENIZER_DATA;
lastCR = false;

22
parser/html/nsHtml5Tokenizer.h
View File

@ -162,21 +162,22 @@ class nsHtml5Tokenizer
charRefBuf[charRefBufLen++] = c;
}
inline void clearCharRefBufAndAppend(char16_t c)
{
charRefBuf[0] = c;
charRefBufLen = 1;
}
void emitOrAppendCharRefBuf(int32_t returnState);
inline void clearStrBufAndAppend(char16_t c)
inline void clearStrBufAfterUse()
{
strBuf[0] = c;
strBufLen = 1;
strBufLen = 0;
}
inline void clearStrBuf()
inline void clearStrBufBeforeUse()
{
MOZ_ASSERT(!strBufLen, "strBufLen not reset after previous use!");
strBufLen = 0;
}
inline void clearStrBufAfterOneHyphen()
{
MOZ_ASSERT(strBufLen == 1, "strBufLen length not one!");
MOZ_ASSERT(strBuf[0] == '-', "strBuf does not start with a hyphen!");
strBufLen = 0;
}
@ -211,6 +212,7 @@ class nsHtml5Tokenizer
inline void appendCharRefBufToStrBuf()
{
appendStrBuf(charRefBuf, 0, charRefBufLen);
charRefBufLen = 0;
}
void emitComment(int32_t provisionalHyphens, int32_t pos);